Mentoring to reduce antisocial behaviour in childhood
BMJ 2004; 328 doi: https://doi.org/10.1136/bmj.328.7438.512 (Published 26 February 2004) Cite this as: BMJ 2004;328:512Data supplement
Posted as supplied by author
Table of studies
Citation Study Group Study type Outcome Key results Comments Cavell TA, Hughes JN. Secondary prevention as context for assessing change processes in aggressive children. J School Psychol 200;38:199-235. 62 children in grades 2 and 3 (USA), average age 7.5, identified by teachers as aggressive using standardised scale. Randomised controlled trial comparing directive mentoring (Prime Time) with non-directive mentoring (Standard Mentoring). Level of aggression and acceptance by others. No significant differences between groups in relation to hostile attributions. Prime Time scored higher on mentoring relationships than Standard Mentoring.
More positive ratings by mentors were associated with lower scores on teacher-rated aggression one year on.
No non-treatment group. Five children lost to follow-up and do not appear to have been included in the final analyses.
Davidson WS, Redner R, Blakely CH, Mitchell CM, Emshoff JG. Diversion of juvenile offenders: an experimental comparison. J Consult Clin Psychol 1987;55 (1):68-75. 213 young offenders, mean age 14.2 years, 83% male, 26% from minority ethnic groups. Average arrest rate in year prior to referral 1.46. Excluded those charged with a serious person crime or on probation.
Randomised controlled trial using six different treatment categories. Police contacts, seriousness of alleged police offences, disposition of police offences, court petitions, seriousness of alleged court petitions, disposition of court petitions, and incarceration. Directive mentoring had significantly lower recidivism rates than the controls. In the two-year follow-up 62% (n=76) and 67% (n=12) of directive mentoring groups had no court petitions compared with 38% (n=60) in the control group.
The treatment groups varied in size, from n=76 to n=12. 7 young people dropped out, and do not seem to have been included in the final analysis.
DuBois DL, Holloway BE, Valentine JC, Cooper H. Effectiveness of mentoring programs for youth: A meta-analytic review. Am J Community Psychol 2002;30:157-97. Systematic review and meta-analysis of 54 studies using mentoring for young people (mean age 19), constituting 60 different samples. Total number of participants/ participants per trial, were not reported. Mixed studies with comparison groups, randomised control groups and before-after comparisons. Both fixed and random effects analyses were conducted. Problem/high risk behaviour Academic/
educational
career/
employment, emotional/
psychological, and social competence.
For fixed effects, modest positive effects for all categories of outcomes. For random effects, effect sizes significant only for problem/high risk behaviour and academic/ educational, and career/
employment outcomes. Of 59 samples analysed, 51 showed effects in a positive direction, one equal to zero and 7 negative.
Results varied significantly according to programme utilisation and risk status of participants. More immediate assessment was associated with greater effect sizes. For random effects analysis, effect sizes based on reports from parents, teachers, or administrative records were non-significant.
Grossman JB, Rhodes JE. The test of time: predictors and effects of duration in youth mentoring programs. Am J Community Psychol 2002;30:199-206. Additional analysis of study by Tierney & Grossman 1998 (see below) using the same sample. Randomised control trial comparing mentored group with waiting list control. Factors associated with duration of mentoring relationships. Relationship duration as effect modifier in outcomes listed below.
Significant, positive effects only observed in relationships lasting 12 months or more. Significant decline in measures of self-worth and scholastic self-perception in relationships lasting < 3 months and a significant increase in alcohol use in those lasting < 6 months.
"Matching" criteria (gender, ethnicity and common interests) were examined as predictors of relationship length. Matching moderately associated with longer relationships. We do not know how "unmatched" dyads occurred. O’Donnell CR, Lydgate T, Fo WSO. The Buddy System: review and follow-up. Child Behavior Therapy 1979;1:161-9. 553 11-17 year olds with behaviour management problems at home, school or in the community. Randomised controlled trial comparing directive mentoring group (Buddy System) with no treatment. Arrest records one year before, during and three years after the intervention. Intervention most effective for those arrested for major offences in the previous year. Young people without previous arrests receiving a Buddy more likely to commit a major offence than those in the control group. St James-Roberts I, Samlal Singh C. Can mentors help primary school children with behaviour problems? Thomas Coram Research Unit. London: The Home Office; 2001. 25 6-10 year olds assessed as above the UK 80th centile for behaviour problems; recruited from schools in areas meeting UK criteria for social and economic deprivation. Matched and controlled before-after study. Not randomised. Scores on Goodman’s Strengths and Difficulties Questionnaire. No significant differences between the mentored and comparison children on any of the measures. Part of a larger evaluation which included interviews with parents, mentors and project workers. 32 children completed the mentoring project; before-after data available for 25.
Tarling R, Burrows J, Clarke A. Dalston youth project part II (11-14) an evaluation. London: The Home Office; 2001. 80 young people aged 11-14 perceived as at risk of not completing education and being involved in offending. Before-after study. Literacy and numeracy scores, school attainment and attendance, number and type of offences committed. Small differences in offending and educational attainment, which did not reach statistical significance. Those who dropped out used as a ‘control’ group. Tierney JP, Grossman JB, Resch NL. Making a difference. An impact study of Big Brothers Big Sisters. Philadelphia: Public/ Private Ventures; 2000.
1,138 young people aged 10-16 years referred by local welfare agencies in the USA. Randomised control trial comparing mentored group with waiting list control. Drug use, alcohol use, theft, property damage, violence towards others, school grades, truancy, relationship with parent(s), self-concept, social and cultural enrichment. Over the 18-month trial young people in mentoring group less likely to start using illegal drugs or alcohol, hit someone or skip school; better relationship with parents and more confident about school performance. Outcome assessment was not blind to trial group. Significant changes seen only in youth-reported outcomes, thus improvements rely on the accuracy of these reports.
Related articles
- Editorial Published: 26 February 2004; BMJ 328 doi:10.1136/bmj.328.7438.473
See more
- An adolescent with disabling abdominal painBMJ December 07, 2016, 355 i6101; DOI: https://doi.org/10.1136/bmj.i6101
- Maternal deaths from suicide must be tackled, say expertsBMJ December 07, 2016, 355 i6585; DOI: https://doi.org/10.1136/bmj.i6585
- Andrew Wakefield calls Trump “on our side” over vaccines after meetingBMJ December 05, 2016, 355 i6545; DOI: https://doi.org/10.1136/bmj.i6545
- Electronic fetal monitoring, cerebral palsy, and caesarean section: assumptions versus evidenceBMJ December 01, 2016, 355 i6405; DOI: https://doi.org/10.1136/bmj.i6405
- Clinical commissioning groups will be rated on sepsis careBMJ November 25, 2016, 355 i6361; DOI: https://doi.org/10.1136/bmj.i6361